Skip to content
TablePage.ai Open the app

The DASH Spreadsheet Error Wasn’t Just a Bad Entry

Use the DASH overrun as a case study for controlling placeholders, reconciling actuals and publishing a reviewable budget table.

Share X in f
Wei Hu

A $3.2 million hard-coded entry in a budget workbook contributed to a much broader reporting failure at Alexandria’s DASH bus system. The practical lesson is not merely “check your spreadsheet.” A budget tracker needs controls that make provisional values visible, reconcile the tracker with the accounting system and prevent an unreviewed file from becoming the accepted version of reality.

What happened at DASH

DASH estimated that it finished fiscal 2026 with a $3.7 million unfavorable variance: approximately $2.97 million in operating costs and $700,000 in capital spending. Its September 9 board packet identifies a $3.2 million hard-coded entry error that overstated the projected year-end position and was the primary reason the workbook continued to suggest a near-balanced result. The figures were still subject to the City of Alexandria’s final close when the packet was issued (DASH board packet, September 9, 2026).

The Alexandria Brief described the value as a spreadsheet “placeholder”. The official account is more precise: it calls it a hard-coded entry error. That distinction matters. A placeholder can be legitimate while data is pending; a hard-coded amount becomes dangerous when the workbook does not identify its status, source, owner and expiry date.

The workbook was not the only problem. DASH reported that its tracker was not consistently refreshed or formally reconciled to the City’s Munis accounting records. The main forecasting workbook lacked consistent secondary review, two relevant funds were not combined in one recurring view, outstanding commitments were not always visible, and there was no consistent rolling forecast or escalation process (DASH board packet).

The consequences extended beyond a corrected total. DASH later said pausing outsourced repairs to contain costs was reducing fleet availability and increasing missed trips (The Alexandria Brief, September 23, 2026).

Make every provisional input reviewable

Do not put an unexplained provisional amount directly into a forecast output. Store each input as a record with enough metadata to review it. For example:

period account budget actual commitments remaining_forecast input_status source owner valid_through approved_by
2026-05 Maintenance services 125000 141000 38000 22000 provisional Accounting extract + open POs Finance 2026-06-30 Pending

Illustrative structure and values—not DASH data.

Use a controlled input_status list such as actual, estimated, provisional and approved_adjustment. Excel data validation can restrict a cell to a list and show an error for an invalid entry (Microsoft Support). That keeps labels consistent, but it does not prove that a value was reviewed. Enforce separate rules requiring a source, owner and validity date for provisional inputs, plus an approver for adjustments.

Calculate outputs instead of typing them:

forecast_spend = actual + commitments + remaining_forecast
overrun = forecast_spend - budget

With this convention, a positive overrun means forecast spending exceeds budget. If your organization uses the opposite sign, name and document the field accordingly.

Keep manual adjustments in their own column or table with a reason, timestamp and approver. This makes them searchable and prevents a number typed over a formula from blending into ordinary results.

Add checks that can stop a release

Before anyone exports or publishes the budget table, require a documented pass for these checks:

  1. Source reconciliation: Tracker actuals equal accounting-system actuals for every period and fund.
  2. Fund completeness: Every required fund appears in the consolidated view.
  3. Commitment coverage: Open purchase orders and incurred costs are included or explicitly excluded with a reason.
  4. Provisional-value review: No expired or unapproved provisional value reaches the released forecast.
  5. Formula integrity: Forecast columns contain formulas throughout the expected calculation range.
  6. Variance escalation: Differences above the organization’s defined dollar or percentage threshold have an owner and documented action.
  7. Independent review: Someone other than the preparer signs off on formulas, sources, assumptions and manual entries.

In Excel for Windows, Home > Find & Select > Go To Special > Constants can locate hard-coded cells in a selected calculation range (Microsoft Support). It will also find legitimate labels and inputs, so restrict the scan to ranges that should contain formulas. For a suspicious total, Trace Precedents displays the cells that supply data to its formula (Microsoft Support).

These checks should be proportionate to the decision at stake. The UK government’s analytical quality-assurance guidance identifies peer review and reconciliation to independent sources as assurance activities. It also distinguishes verification—whether analysis meets its design requirements—from validation—whether it meets users’ needs in the intended environment (The AQuA Book).

Publish the reviewed output, not the operating workbook

For public reporting, export a clean table only after the checks and sign-off are complete. Include fields readers can interpret: period, account or category, approved budget, actual spending, forecast spending, overrun or variance, status, source date and revision date.

Define those fields in a data dictionary. Keep subtotal and grand-total rows out of the observation data so readers do not double-count them; publish totals separately as a summary (how to separate totals from data rows).

Alongside the public table, explain:

  • whether figures are preliminary or final;
  • which funds and periods are included;
  • how variance signs should be interpreted;
  • when the source data was last reconciled; and
  • what changed if the release revises earlier figures.

Do not publish the internal operating workbook if it contains employee, vendor, account or other sensitive information. Create a publication copy containing only reviewed public fields.

DASH’s corrective plan follows the same general control pattern: rebuild the forecast from actual spending, reconcile the tracker to Munis monthly, independently validate formulas and assumptions, conduct monthly variance reviews and complete an independent internal-control review (DASH board packet). The durable response to a budget spreadsheet placeholder error is therefore not a brighter warning color. It is a workflow in which every material number has a source, status, owner, reviewer and reconciliation trail.